Hi. Good morning, everyone. Welcome to day one of our Cantor Healthcare Conference. My name's Olivia Brayer. I'm one of the Senior Biotech Analyst here at Cantor, and really excited about this morning's fireside chat, with the CEO of Tenax. We have Chris Giordano. Chris, thank you so much for joining us.
Thanks for having us.
I had a whole list of questions for you, but I'm going to throw them out because you guys just came out with a lot of slides this morning, a lot of new information, a lot of content, and I'm really excited to go through them with you. Maybe I'll just give you the stage a bit to give us a sense of the data that you've gone through, maybe some of the high-level takes that you guys are seeing from your data set. There's a lot for us to run through.
Okay, great. Thanks. Yeah, a little bit of a surprise. Half an hour before we talked to put what we view as the full analysis of the study out in front of the public. on August 10, we had been unblinded for a handful of days. We saw the result. We needed to get it out there. The information we shared brought up a lot of good questions, and so we aim to answer those now. The ESC late breaker, not quite two weeks ago now, answered a lot of questions, too. Really, it was just before that we received the full stat package. So, we're happy to get it out there and answer a lot of questions. Our view overall is that you can't run from it or avoid it.
The primary endpoint missed statistical significance, but there is an enormous amount of very clear information, and a lot of it is now with the public, that will help guide us to the right patients who respond with an improved walk to levosimendan. There is also, in the data that people will see in much more detail than before, a clear indication that in this very complicated patient population who have volume overload on both sides of the heart, we have a drug that is active as a biventricular drug. The NT-proBNP decrease magnitude is so great that it sort of led us to say, "Why is it greater than PERSIST? Why is it greater than HELP?" And the answer is in the echo data, and basically the overloaded left ventricle that defines HFpEF and the overloaded right ventricle that needs to be targeted in any PH patient.
They are both putting out less NT-proBNP. That is our conclusion. The pulmonary artery pressure is coming down at a very impressive rate. And both of these things happen across the population. So what the full data reveal, of course, is still that patients who do not have much room to improve when they enter the study do not improve much. It also will answer a lot of questions about the safety of the population overall. A lot of misinterpretation was made because of the initial way that we broke that data down by quartiles, and I think the data today shows there is an improvement in every quartile among the patients on levosimendan. There is not some harm being done, and that word just does not apply to our trial.
What looks like a worsening to some is simply the placebo effect in the third and fourth quartiles and a reducing effect of levosimendan on those patients who, again, they just have less room to improve. But across those quartiles or frankly, any of the pre-specified populations, we see NT-proBNP coming down, and we see right ventricular systolic pressure or pulmonary artery pressure coming down in the population. So an ideal drug would address both sides of the heart, and it would provide some support to the right side of the heart. And we also see in the echo data that is out now, I think it is slide 18, that basically at a stat sig level, every parameter among I think 25 or 30 on that page is going the right direction and is not crossing the line of identity. Two of them are not stat sig.
The echo data also supports that the calcium sensitization mechanism of the drug is helping these patients. The NT-proBNP coming down so much is showing that cardiac wall stress, likely on both sides, is coming down. And we see the pulmonary pressure dropping at a level that would get a Group 1 drug approved. So that tells us that in the top-line review, we are right. We need to avoid patients who do not have much room to improve, but that is who we will enroll in LEVEL-2. When you see a result like that, you say, "What happened? The drug did not work," or, "The trial failed.
Well, that was going to be my next question. Can you guys hear me okay? That was going to be my next question. You are clearly encouraged by what you are seeing from this drug. You clearly have conviction that the drug is doing the right thing in this patient population. Ultimately, why do you think the primary endpoint missed, despite some of the reassuring data that we will go through this morning?
Yeah. The drug is working more effectively biologically than we thought it would. The primary endpoint missed because, and there are a couple of good slides to walk through here, a great number of patients who walk too far at baseline were enrolled in the trial. We did not, three-quarters of the way through enrollment, start getting nervous about this and say, "Wait a minute. We have got all these people walking 400 m. Why are there so many 50 and 40-year-olds in our trial?" These were not concerns. Patients with a 10% improvement who walk 450 m walk 45 m further. What you see in our data is that those patients in that fourth quartile on placebo, they walk, I think, 15 m further. It is not a huge improvement, but it is not really good for the primary endpoint.
That is what we learned from the trial, that we really did not appreciate the degree to which a patient's baseline health status would limit their improvement, as shown in the Six-Minute Walk. There have not been trials that instructed us that that would be the case above 400 m. We also know that from HELP, the wedge pressure in patients with really high wedge pressure comes down a lot. Some of them walk a lot further when that happens. So we now know what we did not know during the enrollment, that those patients are simply not going to reveal treatment benefit. Their NT-proBNP comes down, their pressures come down. Their profile in terms of their safety, whether they are above or below the median, is the same.
And that is another thing that is in the data that we put out today that I think was a legitimate reaction to the quartile walk data. The question comes up, well, if you only treat half of that population a second time, are we going to see a lot more of the? And the analysis is there and the answer is no. Treatment emergent adverse events are basically the same across the group. That is very reassuring. We probably should talk about safety as soon as we open our mouths to talk about this study, but it is foremost, but it is very encouraging.
Okay, great. Well, why don't we go through some of the slides in a minute? But I do want to ask, as you think about enrolling too healthy of a patient population, I think you all still have conviction in your LEVEL-2 study. I mean, how do you reconcile all of that, right? Because at the end of the day, LEVEL-2 now is really what matters going forward for your company.
Right. Okay, let me comment on it as I go through, right? How do you reconcile, you missed on primary endpoint. You're really excited about these pre-specified endpoints that are biological. These are important endpoints. You see a big translation to walk, but you still missed, right? Let me jump forward. I have to put this up there. People remember the LS mean is the per SAP way of analyzing the primary endpoint. The mean data is there below it. It shows the treatment difference of 8 m. And then this really is the raw data averaged for investors. What we showed on the 10th was the LS means. You basically see a theoretical result that adjusts for a lot of things with the LS means. The FDA is going to require an LS means.
If you have a big imbalance in SGLT2s in the trial, the LS means result will account for that. If you have missing data, which people will remember, we have five missing walks at 12 weeks. It accounts for that. This is the real data. In each of these groups of around 30 patients, this is the average result. What you see there is what Dr. Shah has described several times as an inverse linear relationship between baseline and treatment effect, right? So you're getting 29 or 26 m of absolute improvement from levosimendan if you're below the 333 median. We're getting a little bit of help from the placebo in that first quartile, right? A - 10 m, that's how you get to around 39 in that group. The placebo walk is a little bit improved in the second quartile.
That population looks an awful lot like the HELP result. You're getting much more benefit from the steady state of 3 mg a day. You're not getting quite as much reduction in the group not on drug, but even that population isn't quite as sick, we'll see in a minute, as the HELP population. What you see on the right is, again, the raw data. So it's clear that the two bars that were below that line when we showed this with the LS means, which we highlighted in bright red as a brilliant aesthetic stroke, that what's happening there is the placebo effect is just overwhelming the drug effect. But I'll take a lot of patients in that third quartile, right? And that's where one thing we need to remember is that the number 333 is a randomly selected median.
You can see that you can push that number higher and still have plenty of treatment effect shown.
So maybe before you, if you do not mind going back to that slide quickly. I think a big question coming out of, and a big question that I had coming out of your initial data set was, are you doing harm to patients, right? In that second half, the healthier patient population, because there was such a difference in terms of the way levosimendan performed versus placebo. I mean, here I am looking at the graph and you never actually see levosimendan underperform in terms of where baseline was, right? You are always improving over baseline. I mean, how do you explain that placebo response in those patients? Because it is a pretty striking placebo response.
I think we have looked a lot at, there is a couple of recent large studies of normal, healthy Six-Minute Walk distance by age that came out in 2025 and 2026. Stuart, as you can imagine, is digging deep into this. Your Six-Minute Walk declines precipitously starting at the age of 60. What we have in our trial that are largely concentrated on the right half of the screen right now are younger patients with lower right ventricular pressures at entry, with less NT-proBNP, and most importantly, we found through not just the pre-specified analyses, not just the normal digging that you do, but by kind of a grand view of every factor that contributes, right? We had a major analysis done by the statistics team in the last couple of weeks. What we see is that those patients just walk too far at baseline. That is the primary indicator.
You do not have that group walking much farther because they are better treated. In fact, the best-treated patients in the trial, which for us is now an indication of a patient whose doctor wants to treat them more, they are on the left side. The best use of state-of-the-art therapy in this HFpEF population tends toward the left side of this. So I think what you see there is, the 25 m is remarkable. The 15 is very. Again, that is not quite a 10% improvement in that patient. That patient in the third quartile who is walking 333 to 384, they are getting 12 m from levo, and they are getting double that from placebo. I think you have basically got to shrink the size of those columns well below 30 to start seeing a negative effect on levosimendan, right?
That to me is one of the things that we really needed to clarify, and that I think, again, smart people seeing limited data started to draw the conclusion that there is worsening going on here. Clinical worsening is an endpoint. It is a secondary endpoint in this trial. It is a defined endpoint, and it did not happen more on our drug than it did on placebo in this trial. It is a misuse of the term. There is no harm being done to these patients. Their walk is showing a decline as they are healthier at baseline. That is really what we see.
As you think about the different cohorts, in terms of the baseline Six-Minute Walk score, are you also seeing these different outsized treatment effects across some of the other secondary endpoints that you looked at?
First secondary endpoint, KCCQ, and I would encourage investors looking at our data to talk to experts from the HFpEF world about this, is going to do us not a whole lot of good. As we have dug into characteristics and looking at characteristics of patients groups against the effect, we see a lot of very encouraging stuff. This does not really happen with KCCQ. We tend to see a small but clinically meaningful change in placebo groups across this trial, and we see, in many cases, a slightly higher suggestion of noise in our patients. It is often a couple points higher, but there is no stat sig. The KCCQ is not going to help us. When you go down from there in our secondaries and you look at clinical worsening and you look at change in functional class, again, the results are not stat sig.
There is just not enough worsening in 12 weeks to see clinical worsening in this population. I do not know in 26 weeks if there will be, but those are aspects of the trial that are under review. We are thinking fast and hard about what we want to do with secondaries. Remember, secondaries are not required for approval. They help you expand your label. They help you make additional claims about the benefit of your drug. We have in the dramatic RVSP lowering and the dramatic NT-proBNP improvement in these patients, along with other things in the data, opportunity to get good secondaries. FDA has to have clinical secondary endpoints.
Mm-hmm. Okay, great. I think the other big question is, now that you do have this full data in hand, as you think about the main overlaps between drug effect and the different patient traits and characteristics that you see, or maybe I'll just let you take us through the data.
Yeah, let me walk you through RVSP and there's more data here now. I think a lot of it should answer that question, hang on, is this the same group of patients? I mentioned more kind of the grand analysis, right? You do a lot of multivariate analyses. They're pre-specified. Then you do a multivariable. I talked about this on the 10th. The work had just gotten started. It's done now. There isn't something in our data that points more clearly and robustly to the patient's health status as reflected in their baseline walk that will help guide us to the right patients. There is information. There's plenty of information. There are other things that we are contemplating doing. But basically, removing patients who walk too far at baseline, as this shows, is really going to make the difference.
This slide compares our phase II data on the right with our phase III data on the left, and you see this greater than 25 m effect twice. You see it in patients who are walking very similarly. On the right, in HELP, the mean was 282. The mean of our below the median population, the bottom half of the trial in terms of baseline walk, is 250, so a little bit lower, and you see a good treatment effect. You see less decline in this group of patients who, even in this analysis, aren't as often a functional class three patient. This basically shows a similar thing. Here to me is why the primary endpoint failed. If you compare the population, which remember in HELP it's much smaller, but you can see the HELP population essentially was advantaged by that little blue rim on the left.
There's a greater density of patients walking toward 100, or even below 100 in that trial. What hurt LEVEL, again, a much larger trial, but that big group who are walking greater than 300 m. That's what explains the results at the bottom. The last bar there is LEVEL overall, but the first two walk cohorts or quartiles in our trial are basically very similar to HELP. That's essentially the challenge that we have got to overcome in LEVEL-2 is just keeping that big blob of yellow patients on the upper right out.
I think Dr. Shah was smart to show this and demonstrate that the CADENCE population, which I think was around 275, both placebo and overall their result when they got 20 m by combining their two doses, very similar to HELP. Basically, we got around 20 m if we put that same group in terms of their baseline walk in. I think what we're seeing here is that it's possible to get a clinically meaningful improvement in Group 2. LEVEL is the first phase III Group 2 trial to even complete enrollment. Yes, the primary endpoint missed, but it's clear that because our median is way over to the right of where we needed it to be.
We have a way forward. When we start looking at RVSP, this is again an analysis of patients if you look at the bottom right. This is an analysis of patients based on how far they walked, again, at baseline. This is the below and the above median group. Look at the reduction in RVSP in the sicker patients. They had higher pressure to begin with, but it comes down at a level that, again, could be associated with a Group 1 drug in terms of reducing pressures on the arterial side. We had a reduction in PA pressure in HELP that was numerical. It was not stat sig. The P value on this is 0.009.
When I say that the drug actually is more efficacious than we expected, it's the NT-proBNP lowering effect of 50% plus the fact that you're getting a robust result of almost 5 m lower. This is just the treatment effect in that group. This is the echo parameters. There's not a bar hitting the line of identity and everything is stat sig except for the last two. This to us is demonstrating in terms of pressure, stress, et cetera, many of the echo parameters that this drug is helping on the right side and it's helping on the left side. It's really important to remember the reason that the Group 2 patient has eluded a lot of great drug developers and a lot of great PAH drugs is that when you go after one side of the heart, you may cause a problem on the other.
If you just want to lower PA pressure, watch out for the lungs, watch out for the right side of the heart. What we see here is that we're lowering BNP in such quantities because we're helping both ventricles. The overload that characterizes both HFpEF and PH is coming down. This is another analysis of patients. I know the GLP use caused a lot of concern, so these are patients who are on GLPs in the middle and then on basically GDMT for HFpEF on the left and the right. Again, this is the below the median group with a 26 m improvement. But if they're on SGLT2s they get another 10 m of improvement. Levosimendan, 3 mg a day, works beautifully with SGLT2 inhibitor prescribed patients. On the right you see MRAs. I think 60% of the patients in the trial are on MRAs.
A lot of them they should be. Again, you see an enhanced treatment effect. Is that because these patients are having a biological concert in the drugs that they are? No, it is because their doctors are trying to do everything they can to help them, we believe, and those are the patients who fortunately are going to be helped the most by levosimendan. What if you, instead of dividing the patient population by walk, you divide them by where is their NT-proBNP at the start of the trial? On the left you see those who are below the median by NT-proBNP. Sorry, above the median. These patients' heart walls are under more stress than those on the right. In this trial, there are patients with NT-proBNP in the 40s, in the 50s, less than 100.
In most HFpEF trials that are trying to enrich the population, you might have an NT-proBNP cutoff of 300. You might have an NT-proBNP cutoff of 600 if you have AFib, because the AFib can push your NT-pro up. What we have is a population that when they are below that median, they are actually walking the wrong direction. Again, those are the healthier patients. Guess what their Six-Minute Walk looked like at baseline much of the time. If I showed you a quartile analysis, or sorry, a scatter plot, you would know very quickly which of the four quarters you want to be in. Here is a group of trials that we list and again this is what gives you conviction. All completed trials mostly other than CADENCE I believe. I am not sure about PRESERVE, but most of these completed before we started.
So painful to look at but this is a list of trials in which in PAH on the left in Group 2 and HFpEF on the right, the first being a DAPA study, you do get a good improvement. There is a 20 m improvement in that population on SGLT2 ones because their baseline we believe is 240. There are SGLT2 trials with much higher baselines that don't give you that LEVEL of benefit. The CADENCE patients who did the best maybe it was the dose they were on, maybe it was the dose of baseline walk they were on. That is what we see. There is a 45 m difference between the two treatment arms in CADENCE. The one that was 45 degrees lower had a 14 m improvement. The one that was higher had a reduction by 6 m, 7 m.
What we are seeing now that we look at many PH trials is that 450 in this group of patients is just way too high a ceiling, and that is why we have lowered it.
Okay. When you say you have lowered it, as you think about LEVEL-2, right? I do not know if you are at this point ready to comment on LEVEL-2 or not, but how do you And I know you said that 333 m number was arbitrary and random, right?
It is randomly selected by the
By the study
in randomization, basically. Yeah.
Right. How do you think about what the right cutoff should be for these patients ultimately?
You can do a couple of things. It is truly inverse and linear. As we go from 330 to 350 to 375 to 400 back to 325, the treatment effect follows. There is a point at which you get zero meters of improvement in this trial. Now if you try to decide, is it 357 or 352? You are now talking about these three patients that you are not going to enroll again in a larger trial. You look across that and you make the right judgment. I would say historically, you get stuck in that project manager's triangle of quality, speed, money. Pick two. We are not doing that here. We are going to take the time we need and involve the number of sites to enroll the patients we need in LEVEL-2, whatever that ultimate number may be.
We are going to make sure, though, that we are not enrolling patients who are sort of, in many cases, the healthiest one in the clinic. That is one thing that we have learned from a number of investigators who have seen the results. They have said that there are a couple of things we can do differently in our protocol, not even in terms of a criteria, but just a requirement for how we conduct the assessment of patients initially to make sure that we are not getting that patient who, of course, they can do an exercise right heart cath. They will be glad to do an exercise right heart cath. That is not your typical frail 75-year-old woman in a HFpEF clinic, especially as we look at patients in other countries.
We think without having to preclude it, we got a much healthier HFpEF patient than you would typically see in a trial.
What is the risk that as you enroll a sicker patient population with a higher disease burden, does safety or adverse events, I mean. Well, you front ran my question.
Yeah, no, this I think is one of the best questions that you and others asked starting kind of on day one, day two. Well, hold on. What we've got here is the treatment emergent adverse event. This is greater than those assigned by the blinded physician to the therapy. It's not treatment-related. It includes those, but it's anything that comes up once they start treatment. In the 12-week period, what you see on levosimendan in three columns on the left is the overall, and then the below median and the above median. On the right, you see placebo.
The comparisons are there, and basically what you have is the type and the rate of these events are typical of a HFpEF population across, and there's really no difference that appears in the frequency between groups, between those whose baseline was below 333 or above or equal to it. I think that was a smart question. We looked at it very quickly. The data's there now. I encourage people to look at it. We talk about it a lot as a more disease-burdened population because there are several things that indicate who are the patients that are going to benefit from levo the most and not have this phenomenal placebo response. But disease burden is just a made-up term. They walk less, they're on more therapies, they're older, their NT-proBNP is higher, and their pressures are higher pretty consistently.
What you'd assume then is you're going to have more patients getting really sick. We don't see it in the 12 weeks.
Okay.
Right.
I know you didn't necessarily have this level of data set at ESC, but you were just at ESC, and your team. What was the feedback that you got from the KOLs there?
We got really encouraging feedback from the HFpEF community. I think the bias among investors is you got to hit on your primary endpoint, and you got to do it on this date, and I'm impatient. The bias in the HFpEF world is very different. Most trials are neutral at best. They look at the data. They look at the biological effects. They can't believe 50% lowering practically of NT-proBNP. They know what it does, right? It's in the deck, and I think Dr. Shah finished on that data for a reason. Here is an association between reduced cardiovascular events, the kind of events like hospitalization and death, all-cause mortality, cardiovascular death. These are the things that get measured in a large, long trial in drugs that treat heart failure. This is the association between their level of NT-proBNP lowering and their mortality.
So bottom left in all caps, this is projected. This is not trial data except for the NT-proBNP. We're basically lowering 50% what active Entresto, MRAs, neprilysin inhibition, et cetera, they get it to about 20%, 27% at best. We're basically double. So where would the mortality be in this patient? A HFpEF doc at ESC who looks at this says, "That looks really great.
Then when you show them, "Well, here's the problem with the Six-Minute Walk," they throw their hands up in the air, and they get frustrated. In Europe, there isn't the same association with EMA that there is with FDA and the Six-Minute Walk. We have a registrational path for this drug to get it into a lot of patients. It will help a lot because of the Six-Minute Walk Test, because of the PH pathway, and we're sticking to that. We're disappointed with not hitting stat sig, but we are not disillusioned.
Right.
We are not disheartened. We are going at it, and that's the plan. In terms of the LEVEL-2 study, I think the protocol design paper will come out at a certain point, and it will reflect all of the updating clinical development thinking about the product.
Yeah.
I would stay tuned for updates on timing, size, budget, et cetera. Obviously, it's going to be a little tougher to enroll a population if you cut it in half by walk. I think it's going to be a little easier to enroll in the 15, 16 countries we've opened, especially after ESC. People knowing they can give them two or three, knowing they now have the oral drug that they've relied on in heart failure to help the left side of the heart for ages. I think there are going to be a lot of physicians who want to put their sicker patients. I think it was a bit of a coup to go to Europe with late-breaking science, supposedly, and to tell them that levosimendan works better in sicker patients. They've known that for decades.
It was not a surprise to them, but I think it's going to help us enroll.
Okay. Well, we're out of time, but it sounds like more to come on LEVEL-2. Any kind of last thoughts just on when we'll start to get some of those updates? It sounds like making sure you're enrolling the sicker patients is really what matters the most.
Yeah. I would say it's not going to be a long list of changes we need to make. Our sites already know that. We had a handful of patients who walked more than 333 m in the trial when we read out the data, and that number is really just not going to change. I think the investors can just rely on that. When the steering committee and the company say, "Take a look at this data, here is an update. These patients are undermining the effect." LEVEL is a test of the Six-Minute Walk in these patients, and the data biologically is very supportive. It's frustrating to see what happens with the walk with those patients. But the investigators are basically ready to go, and they're going to stop enrolling those patients.
Okay. Last question, do you plan to meet with the FDA before coming out with any sort of LEVEL-2 updates?
We'll keep all of those regulatory and sort of core strategic decisions pretty private.
Okay.
We know the FDA likes sensible enrichment, and to date, there have been no concerns.
Okay, great. Thank you very much, Chris, and thank you for sharing this update with us today. Great to talk through it with you.
Sorry you didn't have more time to see it first, but thank you for the opportunity. We appreciate it, Olivia.